A method for screening a tumor pathogenic gene and a storage medium

Abstract

The application relates to a tumor pathogenic gene screening method and a storage medium, and the method comprises the following steps: S1, screening tumor differential expression genes from original transcription profile data, and then identifying a tumor pathogenic related gene subset; S2, coding population individuals by taking the number of gene expression in the tumor pathogenic related gene subset as the chromosome length and taking the transcription number as the gene position; S3, taking the minimum number of gene subsets and the maximum classification accuracy index as optimization targets, establishing a Two-Archive multi-objective optimization model based on information sharing of tumor non-redundant pathogenic genes and initializing; S4, performing limited mating selection between populations, information sharing and information compensation until a termination condition is met, and outputting optimal tumor pathogenic gene screening results; wherein, a child advantage complementary mechanism driven by a boundary solution is used for population updating. Compared with the prior art, the application has the advantages of high efficiency and high screening accuracy.

Description

Technical Field

[0001] This invention relates to the field of gene screening technology, and in particular to a method and storage medium for screening tumor pathogenic genes. Background Technology

[0002] Tumor transcriptional profiling data are generally characterized by "small sample size, high dimensionality, and high noise." Most gene features can have tens of thousands of dimensions; however, from a biological perspective, only a small number of genes are truly relevant to the phenotype of tumor samples. Therefore, it is necessary to select pathogenic genes related to tumor formation and development from tens of thousands of genes. The processing and analysis of high-dimensional tumor transcriptional profiling data requires techniques such as feature selection, dimensionality reduction, and regularization to address the curse of dimensionality. The high noise level of transcriptional profiling data necessitates noise filtering and normalization during analysis to improve data accuracy and interpretability.

[0003] The advantages of commonly used filtering methods for screening tumor-causing genes are that they can quickly remove a large number of irrelevant and noisy genes and are simple to apply. However, their disadvantage is that when selecting tumor genes, filtering methods often only measure the correlation between a single gene and the tumor, without considering the interactions between genes. Since tumorigenesis is the result of the interaction of multiple genes, the resulting subset of genes contains a large number of redundant genes. Redundant genes not only increase the computational cost of classification algorithms but also affect classification accuracy. Identifying the optimal subset of genes from the original gene data—that is, the subset of pathogenic genes that does not contain redundant genes—is a crucial step in tumor gene selection. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of the existing technology and provide a method and storage medium for screening tumor pathogenic genes, which is suitable for small sample, high-dimensional, and high-noise tumor transcription profile data, and can screen tumor pathogenic genes more efficiently and accurately.

[0005] The objective of this invention can be achieved through the following technical solutions:

[0006] According to a first aspect of the present invention, a method for screening tumor pathogenic genes is provided, the method comprising:

[0007] Step S1: Screen out differentially expressed tumor genes from the raw transcription profile data, and then identify a subset of tumor-pathogenic genes;

[0008] Step S2: Encode individuals in the population using the gene expression count in the subset of tumor-related genes as the chromosome length and the transcription number as the gene location;

[0009] Step S3: Using minimizing the number of gene subsets and maximizing the classification accuracy as optimization objectives, establish and initialize a two-archive multi-objective optimization model for non-redundant pathogenic genes of tumors based on information sharing.

[0010] Step S4: Based on the Two-Archive multi-objective optimization model based on information sharing, restrictive mating selection, information sharing, and information compensation are carried out among the populations until the termination condition is met, and the optimal tumor pathogenic gene screening results are output; among them, a boundary solution-driven offspring advantage complementarity mechanism is used for population renewal.

[0011] Preferably, in step S2, binary encoding is used to encode individuals in the population, x i =0 means that the i-th gene was not selected, and equal to 1 means that the gene was selected. Each individual represents a non-redundant combination of gastric cancer genes.

[0012] Preferably, the information-sharing-based Two-Archive multi-objective optimization model divides the non-dominated solution set into a convergence archive (CA) and a diversity archive (DA), which are used to guide the population to converge to the true Pareto front and to increase the diversity of the population in the objective space, respectively. The convergence archive (CA) is a set of individuals with high convergence, in which the individuals represent better quality non-redundant gene combinations for gastric cancer. The diversity archive (DA) is a set of individuals with high diversity, in which the individuals make the population distribution more uniform.

[0013] Preferably, the objective function expression of the Two-Archive multi-objective optimization model based on information sharing is:

[0014]

[0015] Where, x i Let f represent the i-th gene. The objective function f1 calculates the number of gene subsets, and f2 calculates the classification error rate.

[0016] Preferably, the evolutionary process of the information-sharing-based Two-Archive multi-objective optimization model includes:

[0017] 1) Determine the crossover probability pc, mutation probability pm, and restricted mating selection probability δ during the evolutionary process;

[0018] 2) Generate upper and lower bounds for decision variables: Since gene screening is a discrete optimization problem, the decision variables take the values ​​of 1 or 0;

[0019] 3) Randomly initialize the population, encode the tumor genes in binary form, and use 0 or 1 to indicate whether the current gene is selected. Each individual in the population is a combination of non-redundant pathogenic genes of the tumor. At the same time, calculate the target value of each individual in the population; initialize the convergence archive CA and the diversity archive DA to be empty.

[0020] 4) During evolution, a restricted selection mating strategy is used for population reproduction:

[0021] 5) Mutation operations are performed during evolution;

[0022] 6) Update the convergence archive CA and the diversity archive DA using the non-dominated solution set;

[0023] 7) For the updated convergence archive CA and diversity archive DA, a sub-generation advantage complementarity mechanism based on boundary solution is adopted to complement the sub-generation advantages.

[0024] 8) Perform a truncation operation on the convergent archive CA and the diversity archive DA.

[0025] Preferably, the method of employing a restricted selection mating strategy for population reproduction includes:

[0026] A restricted mating probability is set to determine the probability of selecting one individual from each of the two databases as the parent.

[0027] Generate a random number between 0 and 1. If the random number is less than the restricted mating probability, then randomly select an individual from the entire population as the parent. Otherwise, select one individual from the convergence archive CA and the diversity archive DA as the parent of the offspring.

[0028] Offspring individuals are produced by crossover and mutation operations on the parents.

[0029] Preferably, the step of using parents to perform crossover and mutation operations to produce offspring individuals specifically involves:

[0030] The crossover operator determines whether the parent individuals need to be crossoverdated based on a pre-specified crossover probability pc, and then selects two parent individuals for crossover. The crossover operator randomly generates new individuals using probability.

[0031] The mutation probability pm is used to determine whether the parent individual needs to undergo mutation.

[0032] By using crossover and mutation operations, offspring individuals can have different gene combinations than their parents, thus increasing the variety of gastric cancer gene combinations.

[0033] Preferably, the population update using a boundary-solution-driven offspring advantage complementarity mechanism, utilizing the boundary solutions in the archive to maintain diversity in the convergent archive CA and to maintain convergence in the diverse archive DA, specifically includes:

[0034] For each individual in the non-dominated solution set, if it can dominate individuals in the convergence archive CA or the diversity archive DA, add it to the convergence archive CA and delete the dominated individual; for individuals that cannot dominate the convergence archive CA or the diversity archive DA, add them to the diversity archive DA.

[0035] Find the boundary solutions of the convergence archive CA and the diversity archive DA. For each objective dimension, find the individual with the maximum or minimum objective value. Use the boundary solutions to drive evolution to balance the convergence and diversity of the population. Add individuals with better diversity from the convergence archive CA to the diversity archive DA, and add individuals with better convergence from the diversity archive DA to the convergence archive CA.

[0036] Preferably, step 8) involves truncating the convergence archive (CA) and the diversity archive (DA). This truncation operation eliminates some individuals from the population, resulting in individuals that meet the optimization criteria and have higher gene combination quality. Specifically, this includes:

[0037] 81) When the database overflows, only delete individuals from the diversity database (DA);

[0038] 82) Calculate the distance from individuals in the diversity archive DA to the convergence archive CA, and find the individual with the shortest distance;

[0039] 83) Delete the individual with the shortest distance until the database size no longer exceeds the threshold.

[0040] According to a second aspect of the invention, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements any of the methods described herein.

[0041] Compared with the prior art, the present invention has the following advantages:

[0042] 1) This invention uses a two-archive multi-objective optimization model based on information sharing to screen tumor pathogenic genes. Based on the existing two-archive multi-objective optimization model, a boundary solution-driven offspring advantage complementarity mechanism is proposed to improve the problem of tumor gene screening optimization stagnation caused by insufficient diversity of the late-stage convergence archive. This achieves efficient solution of the multi-objective problem of tumor pathogenic gene screening.

[0043] 2) The restricted mating selection strategy proposed in this invention increases the selection pressure on the population by selecting individuals from different databases as parents, thereby promoting population evolution. It is particularly suitable for small-sample tumor transcription data and further improves the accuracy of tumor pathogenic gene screening results. Attached Figure Description

[0044] Figure 1 This is a flowchart of the method of the present invention;

[0045] Figure 2 This is a flowchart illustrating the evolution of the Two-Archive multi-objective optimization model for non-redundant pathogenic genes in tumors based on information sharing, as presented in this invention.

[0046] Figure 3 This is a parallel coordinate graph showing the optimal solution set obtained on the test problem DTLZ1 (6 objectives) in the embodiment.

[0047] Figure 4 This is a parallel coordinate graph showing the optimal solution set obtained on the test problem WFG2 (3 objectives) in the embodiment. Detailed Implementation

[0048] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0049] Example

[0050] like Figure 1 As shown in this embodiment, a method for screening tumor-causing genes is presented and applied to gastric cancer transcriptome data. The method includes:

[0051] Step S1: Screen out differentially expressed tumor genes from the raw transcription profile data, and then identify a subset of tumor-pathogenic genes;

[0052] Step S2: Encode individuals in the population using the gene expression count in the subset of tumor-related genes as the chromosome length and the transcription number as the gene location; wherein, binary encoding is used to encode individuals in the population, x i =0 means that the i-th gene was not selected, and equal to 1 means that the gene was selected. Each individual represents a non-redundant combination of gastric cancer genes.

[0053] Step S3: Using minimizing the number of gene subsets and maximizing the classification accuracy as optimization objectives, establish and initialize a two-archive multi-objective optimization model for non-redundant pathogenic genes of tumors based on information sharing.

[0054] Step S4: Based on the Two-Archive multi-objective optimization model based on information sharing, restrictive mating selection, information sharing, and information compensation are carried out among the populations until the termination condition is met, and the optimal tumor pathogenic gene screening results are output; among them, a boundary solution-driven offspring advantage complementarity mechanism is used for population renewal.

[0055] Next, the method in this embodiment will be described in detail.

[0056] The Two-Archive multi-objective optimization model based on information sharing divides the non-dominated solution set into a convergence archive (CA) and a diversity archive (DA), which are used to guide the population to converge to the true Pareto front and to increase the diversity of the population in the objective space, respectively.

[0057] The objective function expression is:

[0058]

[0059] Where, x i Let f represent the i-th gene. Objective function f1 calculates the number of gene subsets, and objective function f2 calculates the classification error rate. For example... Figure 2 The evolutionary process of the information-sharing-based Two-Archive multi-objective optimization model shown includes:

[0060] 1) Determine the crossover probability pc, mutation probability pm, and restricted mating selection probability δ during the evolutionary process.

[0061] 2) Generate upper and lower bounds for decision variables: Since gene screening is a discrete optimization problem, the decision variables take the values ​​of 1 or 0.

[0062] 3) Randomly initialize the population, encode the gastric cancer gene in binary form, and use 0 or 1 to indicate whether the gene is selected. Each individual in the population is a non-redundant combination of gastric cancer pathogenic genes. At the same time, calculate the target value of each individual in the population. Initialize the convergence archive CA and the diversity archive DA to be empty.

[0063] 4) During evolution, a restricted selection mating strategy is used for population reproduction, specifically:

[0064] Set a restricted mating probability to determine the probability of selecting one individual from each of the two archives as a parent; generate a random number between 0 and 1. If the random number is less than the restricted mating probability, then randomly select an individual from the entire population as a parent; otherwise, select one individual from each of the convergence archive CA and the diversity archive DA as the parent of the offspring.

[0065] The process involves using parental crossover and mutation operations to generate offspring individuals. Specifically, the process is as follows: the crossover probability pc is used to determine whether the parental individuals need to undergo crossover, and then two parental individuals are selected for crossover; the crossover operator randomly generates new individuals using probability; and the mutation probability pm is used to determine whether the parental individuals need to undergo mutation.

[0066] By using crossover and mutation operations, offspring individuals can have different gene combinations than their parents, thus increasing the variety of gastric cancer gene combinations.

[0067] 5) Mutation operations are performed during the evolution process.

[0068] 6) Update the convergence archive CA and the diversity archive DA using the non-dominated solution set.

[0069] 7) For the updated convergence archive CA and diversity archive DA, a boundary-solution-driven offspring advantage complementarity mechanism is adopted for offspring advantage complementarity, specifically:

[0070] Using the boundary solutions in the database, we maintain the diversity of the convergent database CA and the convergence of the diverse database DA. Specifically, gene combinations that are closer to the optimization objective need to be maintained in terms of diversity by increasing the number of different combination types. For gene combinations with high diversity, we need to make them even closer to the optimization objective, including:

[0071] For each individual in the non-dominated solution set, if it can dominate individuals in the convergence archive CA or the diversity archive DA, add it to the convergence archive CA and delete the dominated individual; for individuals that cannot dominate the convergence archive CA or the diversity archive DA, add them to the diversity archive DA.

[0072] Find the boundary solutions of the convergence archive CA and the diversity archive DA. For each objective dimension, find the individual with the maximum or minimum objective value. Use the boundary solutions to drive evolution to balance the convergence and diversity of the population. Add individuals with better diversity from the convergence archive CA to the diversity archive DA, and add individuals with better convergence from the diversity archive DA to the convergence archive CA.

[0073] 8) Perform a truncation operation on the convergence archive CA and the diversity archive DA, specifically:

[0074] 81) When the database overflows, only delete individuals from the diversity database (DA);

[0075] 82) Calculate the Euclidean distance from individuals in the diversity archive DA to the convergence archive CA, and find the individual with the shortest distance;

[0076] 83) Delete the individual with the shortest distance until the database size no longer exceeds the threshold.

[0077] By truncation, some individuals in the population are eliminated, resulting in individuals that meet the optimization criteria and have higher gene combination quality. Generally, the number of pathogenic genes is smaller and the classification accuracy is higher.

[0078] This embodiment performs multi-objective optimization on test problems DTLZ1 (6 objectives) and WFG2 (3 objectives), respectively. The parallel coordinate diagrams of their optimal solution sets are shown below. Figure 3 and Figure 4 As shown.

[0079] The electronic device of this invention includes a central processing unit (CPU), which can perform various appropriate actions and processes according to computer program instructions stored in read-only memory (ROM) or loaded from a storage unit into random access memory (RAM). The RAM may also store various programs and data required for device operation. The CPU, ROM, and RAM are interconnected via a bus. Input / output (I / O) interfaces are also connected to the bus.

[0080] Multiple components in the device are connected to the I / O interface, including: input units such as keyboards and mice; output units such as various types of displays and speakers; storage units such as disks and optical discs; and communication units such as network interface cards (NICs), modems, and wireless transceivers. The communication unit allows the device to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0081] The processing unit executes the various methods and processes described above, such as methods S1 to S4. For example, in some embodiments, methods S1 to S4 may be implemented as computer software programs tangibly contained in a machine-readable medium, such as a storage unit. In some embodiments, part or all of the computer program may be loaded and / or installed on the device via ROM and / or a communication unit. When the computer program is loaded into RAM and executed by the CPU, one or more steps of methods S1 to S4 described above may be performed. Alternatively, in other embodiments, the CPU may be configured to execute methods S1 to S4 by any other suitable means (e.g., by means of firmware).

[0082] The functions described above in this document can be performed at least in part by one or more hardware logic components. For example, exemplary types of hardware logic components that can be used, without limitation, include: field programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload programmable logic devices (CPLDs), and so on.

[0083] The program code used to implement the methods of the present invention can be written in any combination of one or more programming languages. This program code can be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing device, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code can be executed entirely on the machine, partially on the machine, as a standalone software package partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0084] In the context of this invention, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. Machine-readable media can include, but are not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0085] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for screening tumor-causing genes, characterized in that, The method includes: Step S1: Screen out differentially expressed tumor genes from the raw transcription profile data, and then identify a subset of tumor-pathogenic genes; Step S2: Encode individuals in the population using the gene expression count in the subset of tumor-related genes as the chromosome length and the transcription number as the gene location; Step S3: Using minimizing the number of gene subsets and maximizing the classification accuracy as optimization objectives, establish and initialize a two-archive multi-objective optimization model for non-redundant pathogenic genes of tumors based on information sharing. Step S4: Based on the Two-Archive multi-objective optimization model based on information sharing, restrictive mating selection, information sharing, and information compensation are carried out among the populations until the termination condition is met, and the optimal tumor pathogenic gene screening results are output; among them, a boundary solution-driven offspring advantage complementarity mechanism is used for population updating. The population update employs a boundary-solution-driven offspring advantage complementarity mechanism. It utilizes the boundary solutions in the archive to maintain diversity in the convergent archive CA and convergence in the diverse archive DA. Specifically, this includes: For each individual in the non-dominated solution set, if it can dominate individuals in the convergence archive CA or the diversity archive DA, add it to the convergence archive CA and delete the dominated individual; for individuals that cannot dominate the convergence archive CA or the diversity archive DA, add them to the diversity archive DA. Find the boundary solutions of the convergence archive CA and the diversity archive DA. For each objective dimension, find the individual with the maximum or minimum objective value. Use the boundary solutions to drive evolution to balance the convergence and diversity of the population. Add individuals with better diversity from the convergence archive CA to the diversity archive DA, and add individuals with better convergence from the diversity archive DA to the convergence archive CA.

2. The method for screening tumor-causing genes according to claim 1, characterized in that, In step S2, binary encoding is used to encode individuals in the population. =0 means that the i-th gene was not selected, and equal to 1 means that the gene was selected. Each individual represents a non-redundant combination of gastric cancer genes.

3. The method for screening tumor-causing genes according to claim 2, characterized in that, The Two-Archive multi-objective optimization model based on information sharing divides the non-dominated solution set into a convergence archive (CA) and a diversity archive (DA), which are used to guide the population to converge to the true Pareto front and to increase the diversity of the population in the objective space, respectively. The convergence archive CA is a set of individuals with high convergence, in which the individuals represent a better quality non-redundant gene combination for gastric cancer. The diversity archive DA is a set of individuals with high diversity, in which the individuals make the population distribution more even.

4. The method for screening tumor-causing genes according to claim 3, characterized in that, The objective function expression of the Two-Archive multi-objective optimization model based on information sharing is: in, Representing the i-th gene, the objective function is... Calculate the number of gene subsets. Calculate the classification error rate.

5. The method for screening tumor pathogenic genes according to claim 3, characterized in that, The evolutionary process of the information-sharing-based Two-Archive multi-objective optimization model includes: 1) Determine the crossover probability pc, mutation probability pm, and restricted mating selection probability δ during the evolutionary process; 2) Generate upper and lower bounds for decision variables: Since gene screening is a discrete optimization problem, the decision variables take the values ​​of 1 or 0; 3) Randomly initialize the population, encode the tumor genes in binary form, and use 0 or 1 to indicate whether the current gene is selected. Each individual in the population is a combination of non-redundant pathogenic genes of the tumor. At the same time, calculate the target value of each individual in the population; initialize the convergence archive CA and the diversity archive DA to be empty. 4) During evolution, a restricted selection mating strategy is used for population reproduction: 5) Mutation operations are performed during evolution; 6) Update the convergence archive CA and the diversity archive DA using the non-dominated solution set; 7) For the updated convergence archive CA and diversity archive DA, a sub-generation advantage complementarity mechanism based on boundary solution is adopted to complement the sub-generation advantages. 8) Perform a truncation operation on the convergent archive CA and the diversity archive DA.

6. The method for screening tumor-causing genes according to claim 5, characterized in that, The method of using restricted selection mating strategy for population reproduction includes: A restricted mating probability is set to determine the probability of selecting one individual from each of the two databases as the parent. Generate a random number between 0 and 1. If the random number is less than the restricted mating probability, then randomly select an individual from the entire population as the parent. Otherwise, select one individual from the convergence archive CA and the diversity archive DA as the parent of the offspring. Offspring individuals are produced by crossover and mutation operations on the parents.

7. The method for screening tumor-causing genes according to claim 6, characterized in that, The process of using parents to perform crossover and mutation operations to produce offspring individuals specifically involves: Based on a pre-specified crossover probability pc, it is determined whether the parent individuals need to be crossovered, and then two parent objects are selected for crossover; the crossover operator randomly generates new individuals using probability. The mutation probability pm is used to determine whether the parent individual needs to undergo mutation. By using crossover and mutation operations, offspring individuals can have different gene combinations than their parents, thus increasing the variety of gastric cancer gene combinations.

8. The method for screening tumor pathogenic genes according to claim 5, characterized in that, Step 8) involves truncating the convergence archive (CA) and the diversity archive (DA). This truncation process eliminates some individuals from the population, resulting in individuals that meet the optimization criteria and have higher gene pool quality. Specifically, this includes: 81) When the database overflows, only individuals in the diversity database (DA) are deleted; 82) Calculate the distance from individuals in the diversity archive DA to the convergence archive CA, and find the individual with the shortest distance; 83) Delete the individual with the shortest distance until the database size no longer exceeds the threshold.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the method as described in any one of claims 1 to 8.

Images